Now that you can create 2D plots for the 2 principal components. Why do we need to visualize 3D dimensions?
As you already know, the more components are used, the more information is stored from the dataset. It definitely depends on the data you're working with, but statistically, the first 3 components capture data variation the best. You can see this dependence on the graph:

The following code was used to create this graph:

import matplotlib.pyplot as plt
plt.plot(pca.explained_variance_, marker = 'o')
plt.xlabel("Eigenvalue number")
plt.ylabel("Eigenvalue size")
plt.title("Scree Plot")

Just as before, we will consider 2 visualization options with the `matplotlib` and `plotly` libraries.

fig = plt.figure(figsize = (10, 8)) 
ax = fig.add_subplot(projection = '3d')
ax.scatter(X_reduced[:, 0], X_reduced[:, 1], X_reduced[:, 2], c=y)
plt.show()

To visualize 3D components with interactive graphics, we use the function `px.scatter_3d`:

fig = px.scatter_3d(
    X_reduced, x = 0, y = 1, z = 2, color = y.astype(str),
    labels={'0': 'PC 1', '1': 'PC 2', '2': 'PC 3'})
fig.update_traces(marker_size = 5)
fig.show()

Function `fig.update_traces` used to resize the marker. Feel free to experiment with changing colors and other graphics options yourself.
The plot above can be moved, orbited and zoomed in. This can be convenient not only for data exploration, but also for their presentation.

Let's take a look at PCA (Principal Component Analysis) from afar. In this section, you will learn what PCA is, why this data processing method is needed and see the real problems it solves.

Learn about the design of the PCA algorithm. This section will show each stage in detail and simply.

Set up your first PCA model using the Python library and solve the challenge provided for you.

Get some insights with 2D and 3D visualization of PCA results.

It's time to find out how we can explain the results. What do the first, second and third components tell us? What can we use the data for in the future?